Frequent multiple hepatitis C virus infections among injection drug users in a prison setting

Authors

  • Son T. Pham,

    1. School of Biotechnology and Biomolecular Sciences, Faculty of Science, Sydney, Australia
    Search for more papers by this author
  • Rowena A. Bull,

    1. School of Biotechnology and Biomolecular Sciences, Faculty of Science, Sydney, Australia
    2. Inflammation and Infection Research Centre, School of Medical Sciences, Faculty of Medicine, Sydney, Australia
    Search for more papers by this author
  • James M. Bennett,

    1. School of Biotechnology and Biomolecular Sciences, Faculty of Science, Sydney, Australia
    Search for more papers by this author
  • William D. Rawlinson,

    1. School of Biotechnology and Biomolecular Sciences, Faculty of Science, Sydney, Australia
    2. Virology Division, SEALS, Department of Microbiology, Prince of Wales Hospital, Sydney, Australia
    Search for more papers by this author
  • Gregory J. Dore,

    1. National Centre in HIV Epidemiology and Clinical Research, University of New South Wales, Sydney, Australia
    Search for more papers by this author
  • Andrew R. Lloyd,

    1. Inflammation and Infection Research Centre, School of Medical Sciences, Faculty of Medicine, Sydney, Australia
    Search for more papers by this author
  • Peter A. White

    Corresponding author
    1. School of Biotechnology and Biomolecular Sciences, Faculty of Science, Sydney, Australia
    • School of Biotechnology and Biomolecular Sciences, Faculty of Science, University of New South Wales, Sydney 2052, Australia
    Search for more papers by this author
    • fax: (61)-9385-1483


  • Potential conflict of interest: Nothing to report.

Abstract

Recent data indicate that multiple hepatitis C virus (HCV) infections (mixed infection, superinfection, and reinfection) are common among injection drug users (IDUs). In this study, we identified and characterized multiple HCV infection episodes among HCV-seronegative IDU prison inmates (n = 488) enrolled in the Hepatitis C Incidence and Transmission Study cohort. Incident HCV infection with detectable HCV RNA was identified in 87 subjects, 48 of whom completed additional follow-up to screen for reinfection or superinfection. All HCV RNA–detectable samples were tested for multiple infection through a series of specifically designed nested reverse-transcription polymerase chain reaction (nRT-PCR) with sequencing and HCV RNA level measurement. Sequencing revealed that 22 of 87 (25.3%) subjects were infected by two or more viruses. Nine (10.3%) subjects were designated as prevalent cases of incident mixed infection, because two distinct HCV strains were detected at the first viremic time point. Fifteen further cases of multiple HCV infection (superinfection or reinfection) were identified, two of which also showed baseline incident mixed infections. The incidence of new HCV infection (superinfection and reinfection) during follow-up was 40/100 person-years (95% confidence interval, 33-44/100 person-years). Spontaneous clearance of viruses from one subtype and persistence of the other subtype after mixed infection was observed in eight subjects. In these subjects, the virus with higher HCV RNA levels superseded the other. Conclusion: This study comprehensively analyzed frequent multiple HCV infections in a high-risk cohort and provides further insight into infection dynamics and immunity after exposure to variant viral strains. The data presented suggest that HCV RNA levels play an important role in viral competition. (HEPATOLOGY 2010;52:1564-1572)

Hepatitis C virus (HCV) infects 2%-3% of the world's population, or approximately 170 million people.1 Injection drug use is the most common route of transmission, with the prevalence in long-term injection drug users (IDUs) ranging from 64% to 94%.1, 2 The majority of infected individuals develop persistent infection (≈75%)3 and an associated risk of progressive fibrosis, cirrhosis, liver failure, and hepatocellular carcinoma.4

The large degree of HCV genomic variation, the lack of protective immunity generated by HCV infection, and frequent opportunities for re-exposure through ongoing injection behaviors underpin the recognized occurrence of multiple HCV infections.5-10 Multiple infection is classified as either mixed infection (also sometimes referred to as coinfection), superinfection, and/or reinfection (see review by Blackard and Sherman11).

Although multiple infection has been well studied for other viruses, relatively little is known about multiple HCV infection. Primary HCV infection in chimpanzees followed by re-exposure to viruses from either homologous or heterologous HCV strains has been reported to be associated with mild hepatitis and partial immune protection.12, 13 In humans, mixed HCV infection is generally transient, with evidence of replacement with the new strain or persistence of the primary strain.8, 14, 15 The reasons for the transient nature of mixed infection have yet to be elucidated but may relate to a more effective immune response against one virus (in contrast to the other),16 competition between the two viruses (with the fitter strain having an advantage),17 or a combination of these factors.

There are limited data regarding the clinical associations of multiple infections. One cross-sectional study by Fujimura et al.10 of 96 HCV-infected patients with hemophilia reported higher alanine aminotransferase levels reflecting greater hepatocellular injury in nine patients (12%) who had mixed HCV genotypes. Another study by Kao et al.18 observed that mixed infection was more often associated with acute exacerbations during chronic hepatitis C infection than monotypic infection.

The reported prevalence of multiple infection in HCV-infected subjects ranges from 5% in a cohort of patients coinfected with HCV and human immunodeficiency virus19 to 39% in a cohort of IDUs.5 The high prevalence of multiple infection in IDUs and the association with high-risk behavior indicates that ongoing injection and needle sharing following primary infection can lead to subsequent acquisition of new HCV strains.5, 6, 20 Longitudinal studies to estimate the incidence of multiple infection include a small number of case series8-10, 15 as well as prospective5-7, 21 and retrospective22, 23 analyses of stored samples. One of the retrospective studies within an IDU population reported a 1.8-fold higher incidence of reinfection (31/100 person-years; 95% confidence interval [CI] 17-62/100 person-years) compared with naïve infection (17/100 person-years; 95% CI 14-20/100 person years).22 In a recent IDU-based prospective study, the incidence rate of reinfection was 2.5-fold higher than primary infection and was associated with injection risk behavior.7 Both of these studies7, 22 suggested that primary infection failed to elicit protective immune responses against subsequent infection in high-risk subjects; however, contrary findings have been reported,21, 24-26 and the existence of protective immune responses following primary infection remains controversial.

It is likely that these studies underestimated multiple infection rates, because the screening methods used lacked sensitivity for detection of viruses present at low levels (<1% of the population) in a single sample, and screening for different viruses from the same subtype was not undertaken.6, 7, 19, 22 For these reasons, further evaluation of the incidence of multiple infection is required.

In the present study, two nested reverse-transcription polymerase chain reaction (nRT-PCR) assays for the detection of multiple infection were developed and validated. The first assay incorporated a set of HCV subtype-specific nRT-PCR primers that amplified a portion of the core region and was used to detect one or more genotypes in a single serum sample. The second assay targeting the HCV core C terminus, envelope glycoprotein 1 and the hypervariable region 1 of envelope glycoprotein 2 (E1/HVR1) was used to detect subtype and genotype changes in longitudinal samples. Using longitudinally collected samples from a prospective cohort of seronegative and HCV RNA–negative IDU prison inmates,27 the objectives of this study were to evaluate the prevalence of mixed infection at incident HCV infection and the incidence of subsequent multiple infection (superinfection, reinfection, strain switch) during follow-up. The natural history, including HCV displacement, of these multiple infection episodes and viral factors that predicted the outcome of viral competition were examined.

Abbreviations

CI, confidence interval; Ct, threshold cycle; E1, envelope glycoprotein 1; HCV, hepatitis C virus; HITS, Hepatitis C Incidence and Transmission Study; HVR1, hypervariable region 1 of envelope glycoprotein 2; IDU, injection drug user; nRT-PCR, nested reverse-transcription polymerase chain reaction.

Subjects and Methods

Study Population.

The Hepatitis C Incidence and Transmission Study (HITS) is a prospective cohort study of HCV-seronegative/RNA-negative, high-risk IDU prison inmates recruiting in 19 correctional centers in New South Wales, Australia. Details of the study protocol have been reported elsewhere.27 All participants provided written informed consent. The protocol was approved by the institutional review boards of Justice Health and the Department of Corrective Services.

Virological Assessments.

All sera were initially tested using a qualitative HCV RNA detection assay (TMA assay, Versant, Bayer, Australia; lower limit of detection, <10 IU/mL). If detectable, quantification was undertaken (Roche COBAS AmpliPrep/COBAS TaqMan test; limit of detection, 15 IU/mL).

Definitions.

Multiple infection is defined as infection with more than one HCV strain. Multiple infection can be further subclassified into mixed infection (infection by two or more heterologous viruses either simultaneously or within a narrow time period),11 superinfection (detection of infection by an HCV strain distinct from the primary infecting strain in subjects with persistent viremia), and reinfection (infection by a distinct HCV from the primary strain following viral clearance).11 A strain switch was classified when two or more distinct viruses were detected during follow-up but the presence or absence of viremia in the interim period had not been established.

Subtype-Specific Primer Design.

In order to detect multiple HCV genotypes in a single sample, a set of four individual nRT-PCR assays were developed using subtype-specific primers that targeted the HCV subtypes 1a, 1b, 2a, and 3a (these are the most prevalent genotypes in Australia, accounting for 94%-99% of all HCV infections).28, 29 Unique sequences (n = 279) of the entire core region were pooled from GenBank and the HCV Los Alamos databases and aligned using Clustal-W. A universal forward primer (Hep287) was designed for use in the first round amplification (Supporting Information Table 1). The remaining primers (three for each subtype) were designed such that the last three to five nucleotides of the 3′ end of the primer were specific to the targeted subtype (Supporting Information Table 1). Amplicon lengths for each genotype were 289 bp (1a), 237 bp (1b), 354 bp (2a), and 241 bp (3a). The overlapping 5′-3′ end of the core region for all for subtypes was 228 bp in length.

Table 1. Subjects with Multiple Infections
Subject ID NumberGenotypeEstimated Time of Subsequent Infection from Primary Infection (Weeks)Estimated Duration of Mixed Infection (Weeks)Viral Load Quantity Ratio (IU/mL) (1o:2o or 2o:3o)Total Number of VirusesInfection Outcome
PrimarySecondary/Tertiary
  • Abbreviations: C, spontaneous clearance; NA, not applicable; P, persistence of the primary strain; PB, persistence of both strains; R, reinfection; S, strain switch.

  • *

    11% E1/HVR1 sequence divergence.

  • Mixed infection identified in the last time point studied.

  • Identical core sequences.

  • §

    13% E1/HVR1 sequence divergence.

3001042a3a0No follow-up10:12NA
3002271a3a0No follow-up10:12NA
3002301a2a0No follow-up105:12NA
3003741a3a0No follow-up1:1012NA
3001573a3a*0172:12NA
3002892b1b09103:12NA
3003031a3a0192:12C
3002031a2a015105:13C
 3a44No follow-up10:1NA
3003041a3a05104:12PB
 3a31No follow-up104:1
        
3000811a3a8414105:12P
3002593a2a5No follow-up102:12NA
3003062b1a13102:12P
3003471a3a2851:12P
3004991a3a1610:12P
3002233a3a§27NANA3P
 1a7034105:1
3000971a3a115No follow-up104:12NA
        
3000013a1b92NANA2R
3002121a1b79NANA2R then C
3000862b3a146NANA2R
        
3001173a1a26NA102:12S
3001553a3a*10NA1:12S then C
3003571a3a7NANA2S

Real-Time Subtype-Specific nRT-PCR for Detection and Quantification of HCV Mixed Infection.

RNA was extracted from sera using the QIAmp Viral RNA extraction kit (QIAGEN, Hilden, Germany). Mixed infection at a single time point was detected by performing four real-time subtype-specific nRT-PCRs. Reverse-transcription was performed with 5 μL of RNA template added to a 15-μL reaction mix containing 1× VILO reaction mix and 1× Superscript enzyme mix (Invitrogen, Mount Waverley, Australia). Reverse-transcription was performed at 25°C for 10 minutes, then at 42°C for 60 minutes. Complementary DNA (5 μL) was added to 15 μL of first-round reaction mix containing 1× iQ SYBR green supermix (BioRad, Hercules, CA) and 0.5 μM of each primer. Reactions were performed at 94°C for 2 minutes, then 15 cycles of 95°C, 52°C, and 72°C for 1 minute, respectively. Second-round PCR was performed for 40 cycles using 2 μL of first-round product added to 18 μL of PCR reaction mix. Reaction mix and conditions were as described above, except that the annealing temperature was raised to 56°C. For quantification, standard curves were generated from serial 10-fold dilutions of T7 polymerase-transcribed HCV RNA, and threshold cycle (Ct) values obtained were used to determine the number of viral genome copies/mL of serum, which was subsequently converted into IU/mL.30 The detection limit of the real-time subtype-specific nRT-PCRs was 86 IU/mL.

T7 Transcription of Core RNA Standards.

In order to evaluate the sensitivity and specificity of the subtype-specific nRT-PCRs, T7 transcripts of the core region were generated using the MEGAscript kit (Ambion, Austin, TX) from HCV RT-PCR products amplified by nRT-PCR using Hep21b and GV33 for the first round and Hep346 and GV35/GV36 for the second round (Supporting Information Table 1). RNA was treated twice with TURBO DNase for 1 hour and purified using Ambion MEGAclear (Ambion). The RNA was quantified by way of spectrophotometric measurement at 260 nm, and the copy number was calculated.

Validation of the Real-Time Subtype-Specific nRT-PCR to Identify Mixed Infection.

Several experiments were conducted to validate the individual real-time subtype-specific nRT-PCR assays. Using both T7-transcribed RNA and serum-extracted HCV RNA, the four subtype-specific nRT-PCR assays amplified only the specified subtype RNA (i.e., 1a, 1b, 2a, or 3a) with no cross-reactivity detected, even in the presence of 1 × 106 copies of alternate serum-derived HCV subtype RNA/reaction (Supporting Information Fig. 1A). The lower limit of detection of the subtype-specific nRT-PCR was calculated as 1 copy/reaction for each targeted subtype (data not shown) and between 1 and 100 copies/reaction using sera of known subtype and viral load (Supporting Information Fig. 1B). To determine the specificity of the subtype-specific nRT-PCR, T7 transcripts from each subtype were separately mixed with T7 transcripts from the three heterologous subtypes in ratios of 1:1×106 copies per reaction. Ct values for each subtype/subtype ratio were compared with the Ct values for the individual subtypes alone at 1 copy/reaction (Supporting Information Fig. 1C). There were no significant differences (P < 0.001 [one-way analysis of variance]) between the Ct values in the presence or absence of heterologous RNA, even at 1 × 106 copies per reaction (Supporting Information Fig. 1C). These results were reproduced using RNA derived from infected serum (data not shown).

Figure 1.

Nucleotide divergence between the different genetic classifications of HCV using uncorrected nucleotide distances (p-distance). Sequences (n = 541) (657 bp [nucleotides 915-1571, H77, accession number AF009606]) were pooled from GenBank. Frequencies of pairwise distances demonstrating HCV genetic classifications are shown for intrasubtype (1a-1a [white]), intragenotype (1a-1b [gray]), and intergenotype (1a-3a [black]). Pairwise distances (%) within the intrasubtype, intragenotype, and intergenotype levels were 2.4%-13.4%, 20.1%-28.2%, and 28.9%-39.0%, respectively, with the means indicated by dotted lines. Three cases who were infected with two different viruses (ID 300155, ID 300157, and ID 300223) from the same subtype are indicated by black circles (3a in all cases). Pairwise nucleotide comparisons from a longitudinal sampling of subjects within the same cohort infected with a single virus (viral evolution, control group, n = 25) are shown as a box and whiskers plot. The box limits represent the 25th and 75th percentiles, the vertical line within box is the mean (1.1% ± 0.7%), and the whisker ends represent the 1st and 99th percentiles (range, 0.0%-2.4%).

Amplification of E1/HVR1 to Detect Multiple Infection on Longitudinal Samples.

The region encoding the last 171 bp of core, E1, and HVR1 (840 bp [nucleotides 744 to 1583, with reference to HCV strain H77; GenBank accession number AF009606]) was amplified by way of real-time nRT-PCR with the HCV primers described in Supporting Information Table 1 and using reagents and reaction conditions described in Tu et al.31

Strain-Specific E1/HVR1 nRT-PCR for Detection of Mixed Infection with Two Viruses from the Same Subtype.

Where potential secondary infection to a virus from the same subtype (e.g., 3a-3a) was detected through sequencing of longitudinal samples, individual sequence-specific nRT-PCR was performed to determine when each virus was present. The first round was run with E1/HVR1 universal primers GV32/GV33 using the conditions described above. For the second round, sequence-specific primers were designed based on the two E1/HVR1 subtype sequences detected.

Sequence Analysis to Identify Multiple Infection in Longitudinal Samples.

Sequencing reactions were performed as described.31 In order to detect HCV superinfection and reinfection (or strain switch where the former could not be differentiated), E1/HVR1 and/or core sequences were generated from samples collected longitudinally, and the pairwise sequence divergence was calculated using the p-distance algorithm. Reinfection or superinfection from a heterologous HCV subtype was confirmed by way of phylogenetic analysis of the core region.31

Thresholds for designation of superinfection and reinfection (rather than viral evolution) were defined on the basis of E1/HVR1 pairwise sequence comparisons using 541 E1/HVR1 sequences from the GenBank database (Fig. 1). Reinfection or superinfection with a different virus from the same subtype (e.g., 1a-1a) was designated when the nucleotide divergence was above the minimum value plus 3× SD (6.0%) for viruses within each subtype (range, 2.4%-13.4%; mean ± SD, 8.7% ± 1.2%) (Fig. 1). Sequence divergence of viruses from different subtypes (e.g., 1a-1b) ranged from 20.1% to 28.2% (mean ± SD, 23.4% ± 1.0%) and of viruses from different genotypes (e.g., 1a-3a) ranged from 28.9% to 39.0%; (mean ± SD, 32.6% ± 1.5%) (Fig. 1).

Epidemiological Analysis.

The estimated duration of mixed infection was calculated using the midpoint between the initial mixed incident infection time point and the subsequent resolution to a single HCV strain. The estimated time to infection with a second virus following a primary infection was calculated as the interval between incident HCV detection and initial detection of the multiple infection (mixed superinfection, reinfection, or strain switch). The incidence of multiple infection was calculated as the person-years rate of new infections with all subjects contributing follow-up time from initial incident HCV detection and censored at last HCV RNA time point. Subjects were not censored at detection of multiple infection, because further cases of multiple infection within individual subjects were possible.

Results

Cohort Characteristics.

Analysis of the first 488 previously anti-HCV antibody–seronegative subjects enrolled in the HITS cohort indicated that the population was predominantly male (65%), with high rates of prior imprisonment (72%) and longstanding injection drug use (mean 8.5 years). During a mean follow-up of 38 ± 33 weeks, a total of 90 incident HCV infections were detected, including 87 (96.7%) subjects with detectable HCV RNA sequences at initial infection. Of these 87 subjects, 48 completed at least one further longitudinal time point following detection of incident HCV infection (Fig. 2).

Figure 2.

HCV multiple infection among HCV incidence HITS subjects. Cases are shown chronologically from baseline screening to follow-up testing. Nine cases of incident mixed infection were identified (10.3%) (oval). Fifteen cases (31.3 %) of superinfection/reinfection were detected (diamonds), including nine superinfections, three reinfections, and three that demonstrated strain switches.

Detection of Incident Mixed Infection.

Eighty-seven incident HCV infection cases who had viral sequences available were analyzed for multiple infection, with an average of 16 ± 28 weeks since the last undetectable HCV RNA sample (range, 0-127 weeks). Nine of 87 (10.3%) subjects were designated as incident cases of mixed infection, because two distinct HCV strains were detected at the first HCV RNA–detectable time point (Figs. 2 and 3). These observed mixed incident infections included 1a-3a (n = 4), 1a-2a (n = 2), 2a-3a (n = 1), 1b-2b (n = 1), and 3a-3a (11.0% divergence) (n = 1) (Fig. 3A).

Figure 3.

Twenty-two subjects with multiple infection. All individuals are shown relative to a time scale that comprised the time of incidence of HCV infection to 42 months (42M). HCV RNA–negative subjects enrolled in HITS-p (n = 87) who became RNA-positive were screened for multiple infection with two or more HCVs. Subtypes are color-coded as shown in the key. Circle sizes correlate to log of viral load (IU/mL) of each HCV population at each time point. Boxes indicate RNA (negative) time points. The 22 cases of multiple infection are grouped into (A) incident mixed infection (n = 9), (B) superinfection with no incident mixed infection (n = 7), (C) reinfection (n = 3), and (D) strain switch (n = 3).

Detection of Superinfection, Reinfection, and Strain Switch.

Core and/or E1/HVR1 sequences were generated for all follow-up time points available for 48 of 87 subjects (mean sampling interval, 26 ± 29 weeks) (two time points, n = 18; three time points, n = 10; four time points, n = 12; five or more time points, n = 8).

Fifteen of the 48 subjects became infected with a new HCV strain during follow-up (cumulative prevalence of subsequent infection, 31.3%) (Figs. 2 and 3). The incidence of new infections during follow-up was 40/100 person-years (95% CI, 33-44/100 person-years). Among these 15 subjects, nine cases of superinfection were identified, including two of the original nine incident mixed infection cases infected by a third virus (1a/2a-1a/3a, ID 300203 and 1a/3a1-1a/3a2, ID300304) (Fig. 3A) and seven other superinfection cases (1a-3a, n = 5; 1a-2b, n = 1; 2a-3a, n = 1) (Fig. 3B). Three cases of reinfection were observed with clearance of a 1a virus and reinfection with a 1b virus (ID 300212), clearance of a 3a virus with subsequent reinfection with a 1b virus (ID 300001), and clearance of a 2b virus followed by reinfection with a 3a virus (ID 300086) (Fig. 3C). In the three remaining multiple infection cases, the incident HCV strain was eliminated, followed by detection of a second HCV strain (3a-3a [11.0% divergence], ID 300155; 3a-1a, ID300117; 1a-3a [ID300357]). In these three cases, we could not differentiate reinfection from superinfection, because no HCV-negative time points were observed between the sampling time points, and therefore designated these cases as strain switch (Fig. 3D). Overall, the cumulative prevalence of multiple infection in this cohort of incident HCV infection was 25.3% (22 of 87 subjects).

Natural History of Multiple Infection.

In 11 of the 22 subjects with multiple infection (Fig. 3A-D), longitudinal follow-up postdetection of mixed infection or subsequent infection were available to investigate the natural history. Persistent infection by a 1a virus with elimination of a 3a virus was observed in three subjects (ID 300081, ID 300304, and ID 300499), with subsequent superinfection of the same virus for ID 300304. Persistence of a 3a virus with clearance of a 1a virus was detected in one subject (ID 300347), persistence of a 2b virus with clearance of a 1a virus was detected in one subject (ID 300306), and persistence of one of two 3a viruses was observed in one subject (ID 300157) (Fig. 3). One subject cleared both viruses in a 1a-3a mixed incident infection within 12 months (ID 300303), whereas another subject with 1a-2a mixed incident infection cleared the 2a infection and was subsequently superinfected with a 3a virus (ID 300203). For the remaining three subjects, one cleared the first and third infecting viruses (3a and 1a) and remained persistently infected with the second 3a virus (divergence between the 3a viruses, 13.0%) (ID 300223); one reinfection subject had spontaneous clearance following subsequent infection during follow-up (ID 300212); and there was one case of strain switch (ID 300155) (Fig. 3). The rate of spontaneous clearance of all viruses in multiple infection was calculated as 19/100 person-years (95% CI, 15-26/100 person-years).

Among 15 subjects with mixed infection, 10 had longitudinal follow-up postdetection of mixed infection (superinfection, n = 5; incident mixed infection, n = 3; both, n = 2). To determine if HCV RNA levels could be used as a predictor of the outcome of viral competition, quantitative subtype-specific nRT-PCR was used. In two of 10 subjects, both viruses were cleared after the first detectable mixed infection (ID 300303), and one maintained mixed infection after 8 weeks before being lost to follow-up (ID 300289). In the remaining eight cases, the virus with the higher HCV RNA level superseded the other virus (Fig. 3, Table 1). Indeed, the HCV RNA level was higher in the primary infecting strain that superseded the incoming strain in five of seven superinfection cases (Fig. 3B).

Mixed infection was generally transient. The longest duration of mixed infection was estimated to be 34 weeks (ID 300223) (Table 1). The mean duration of mixed infection was 13 ± 9 weeks (range, 3-34 weeks) (Table 1). The mean estimated time of infection with a second virus following a primary infection was 48 ± 45 weeks (range, 1-146 weeks; n = 16).

Discussion

Through detailed virological characterization in a prospective cohort using specifically designed molecular methods, we have provided new insight into the burden and natural history of multiple infections among high-risk individuals in a prison setting. Our findings indicate that multiple infections are common and generally transient and that viral clearance was related to a lower HCV RNA level between the competing individual strains.

Existing methods for assessment of HCV multiple infections are either capable of detecting more than one HCV genotype at a single time point or analyze sequences longitudinally to detect reinfection and/or superinfection; however, few studies have combined these approaches. Published methodologies include serotyping14 or RT-PCR–based approaches with downstream processing, including commercially available line probe genotyping assays,23 sequencing of amplicons,6, 7, 19, 21, 22 cloning with sequencing,5 and heteroduplex mobility analysis.32 All of these methods are either limited by the sensitivity of detection of mixed infection as they could only detect strains circulating at relatively high proportions within the quasispecies (1%-10% of the population)5, 6, 15, 19 or could not differentiate between reinfecting/superinfecting viruses from the same subtype.7, 14, 22 They are therefore likely to underestimate the true level of multiple infection. In addition, many of these studies are further constrained by short study periods,14 long sampling intervals,5, 6, 22 and small sample sizes.15 In the current study, the use of sensitive molecular methods to detect low levels of a minor viral population (1 in 1 × 106 genome copies/reaction) and the ability to differentiate between different viruses of the same subtype increased the likelihood of detecting multiple infection. Indeed, the performance of the four nRT-PCR assays used was assessed by sequencing of the amplicons. Assay and sequence results from samples containing either HCV 1a (n = 44), 1b (n = 6), 2a (n = 5), or 3a (n = 60) were entirely concordant, indicating 100% sensitivity and specificity for all subtypes.

A high cumulative prevalence (24.4%) of multiple infection was detected within this IDU prisoner cohort. This rate is comparable with other longitudinal studies of high-risk IDUs that reported prevalences from 20% to 39%5-7, 22 and higher than previous cross-sectional studies among patients with chronic HCV infection that reported mixed infection prevalences ranging from 1.4% to 13.5%.23, 32 In the present cohort, the incidence of new infection during follow-up was calculated to be 40/100 person-years (95% CI, 33-44/100 person-years), which is concordant with data from other seroconverter cohorts of young IDUs (31/100 and 47/100 person-years)7, 22 and higher than the reported incidence of naïve infection (16/1007 and 17/10022 person-years). This finding, taken together with the findings of other studies, demonstrates that multiple HCV infections in a high-risk cohort are common.

The reported incidence of reinfection/superinfection is comparable or higher than the rate of primary infection,5-7, 22 which indicates a lack of significant sterilizing immunity following primary infection. However, these studies were either retrospective22 or lacked a comprehensive analysis of the natural history of multiple infection,5-7, 22 including levels of competing viremia. They also lacked subsequent follow-up once multiple infection was detected to determine the duration of infection or the outcome of viral competition. Therefore, levels of protective immunity could not be assessed. In a recent study by Osburn et al.,21 a reduction in the magnitude and duration of viremia in cases of reinfection was observed, suggesting that adaptive immunity may protect against chronic disease.

Limited data are available regarding the natural history of mixed infection and superinfection in untreated incident cases of HCV infection. Multiple infections were found to be transient in nature in the present study, consistent with previous reports.8, 14, 15 Clearance of one or more viruses following multiple infection was frequently documented in the present report (n = 11), with the rate of viral clearance measured at 19/100 person-years. Indeed, spontaneous clearance of two or more viruses was also observed in three subjects with multiple infection. Clearance of an HCV infection may be triggered when the second strain boosts cross-strain immunity elicited in association with the first infecting strain. Although such immunity has not been examined directly using immunological assays, this outcome is consistent with three studies in which eradication of the primary strain followed superinfection.15, 18, 23

Although host immunity may play a role in determining which virus survives in the setting of transient mixed infection, viral factors may also be an important consideration. In the present study, HCV RNA levels were shown to be a major factor influencing the outcome of mixed infection. In all eight cases where one strain superseded the other, it was the strain with the higher starting HCV RNA level that became the sole infecting virus. There are two hypotheses for this observation. First, the HCV RNA levels could directly relate to the replication efficiency and act as a tool to compete against another strain and evade the immune system.17 Second, the HCV RNA levels may relate to magnitude of the host response; that is, the virus with the higher level may be the strain with lower immune pressure.16 The latter hypothesis was further supported by the observation that the persistent HCV virus following superinfection was the primary strain that may have already evaded the immune response and established persistent infection. Accordingly, further studies with longitudinally collected samples specifically examining cellular immune responses against the initial viral strain and a subsequent reinfection strain are needed to formally examine the evidence for cross-strain immunity and potential protection against chronic infection.

The present study has some limitations. First, the six monthly sampling intervals in the HITS cohort may have allowed additional viremic episodes to be missed if they resulted in prompt clearance, resulting in an underestimation of the frequency of mixed infection. In this regard, it is noteworthy that the mean duration of mixed infection in the cases reported here was 13 weeks. On this basis, a sampling interval of 3 months would be both preferable to improve sensitivity of detection, and feasible in the field. Additionally, it should be noted that the assay systems used in this study could not detect mixed infection involving rare genotypes (e.g., 4, 5, 6), though these genotypes are uncommon in Australia.

This study provides new further insights into the prevalence and natural history of multiple HCV infections. Analyses of the behavioral risk factors for multiple infection are warranted, and immunological studies to examine cross-strain immunity are also needed. Ongoing recruitment and follow-up in the HITS cohort will increase the sample size and follow-up of monoinfection and multiple infection episodes to further resolve the impact on clearance rates and on the disease course of those with chronic mixed infection.

Acknowledgements

We thank Aileen Oon and Brendan Jacka for technical assistance and Suzy Teutsch, Hui Li, and Luke McCredie for assistance in data management and specimen handling. The ongoing cooperation of the prisoners who volunteered to participate in the HITS cohort is gratefully acknowledged.

Ancillary